Analyzing Census Data in R

Ari Lamstein (@AriLamstein)
San Francisco R-Ladies Meetup, May 2015

Introduce yourself to your neighbor

  1. What state are you from?
  2. What county are you from?
  3. What ZIP code are you from?

Today we will use R to understand the demographics of the

  1. State you are from
  2. County you are from
  3. ZIP code you are from


Then we will

  1. Analyze historic data
  2. Look at more variables

The Data

  1. American Community Survey (ACS)
    • Not the decennial census
  2. Very detailed / personal questions
    • “How long is your commute?”
    • “What is your income?”
  3. $400B / year is allocated based on the results
    • Where should money for schools go?
  4. R's acs package
    • Connects to the Census API Directly
    • Deals with confidence intervals, etc.

Part 1

States

library(choroplethr)
data(df_pop_state)

?df_pop_state
head(df_pop_state)

      region    value
1    alabama  4777326
2     alaska   711139
3    arizona  6410979
4   arkansas  2916372
5 california 37325068
6   colorado  5042853

Exercise: What is the population of your home state? Tell your neighbor.

My Answer

df_pop_state[df_pop_state$region == "new york", ]

     region    value
33 new york 19398125

Boxplot

boxplot(df_pop_state$value)

plot of chunk unnamed-chunk-3

Exercise: Draw one interpretation from this plot

Choropleth Map 

state_choropleth(df_pop_state)

plot of chunk unnamed-chunk-4

Exercise: Draw one interpretation from this map

Exploratory Data Analysis 

state_choropleth(df_pop_state, num_colors = 2)

plot of chunk unnamed-chunk-5

Exercise: Draw one interpretation from this map

Continuous Scale 

state_choropleth(df_pop_state, num_colors = 1)

plot of chunk unnamed-chunk-6

Exercise: Draw one interpretation from this map

More than population 

data(df_state_demographics)
?df_state_demographics

colnames(df_state_demographics)

[1] "region"            "total_population"  "percent_white"    
[4] "percent_black"     "percent_asian"     "percent_hispanic" 
[7] "per_capita_income" "median_rent"       "median_age"

Exercise: Make a choropleth map of one statistic above. Share it with your neighbor.

My answer 

df_state_demographics$value = df_state_demographics$percent_white

state_choropleth(df_state_demographics, num_colors=2)

plot of chunk unnamed-chunk-8

Part 2

Counties

County Population

library(choroplethr)
data(df_pop_county)
head(df_pop_county)

  region  value
1   1001  54590
2   1003 183226
3   1005  27469
4   1007  22769
5   1009  57466
6   1011  10779
  • (region, value) pairs again
  • County FIPS code

County FIPS codes

Exercise: Use google to find the FIPS code of the county you are from. Share it with your neighbor.

Exercise: What is the population of the county you are from? Share it with your neighbor.

Write down your county FIPS code. We'll be using it later.

My Answer

# 36059 is the FIPS code for Nassau County, NY
df_pop_county[df_pop_county$region == 36059, ]

     region   value
1858  36059 1338712

Boxplot

boxplot(df_pop_county$value)

plot of chunk unnamed-chunk-11

Exercise: Draw one interpretation from this boxplot

Choropleth Map

county_choropleth(df_pop_county)

plot of chunk unnamed-chunk-12

Exercise: Draw one interpretation from this map

county_choropleth(df_pop_county, num_colors=1)

plot of chunk unnamed-chunk-13

Exercise: Draw one interpretation from this map

Zoom

county_choropleth(df_pop_county, state_zoom="california", num_colors=4)

plot of chunk unnamed-chunk-14

Exercise: Draw one interpretation from this map

More Demographics

data(df_county_demographics)
colnames(df_county_demographics)

[1] "region"            "total_population"  "percent_white"    
[4] "percent_black"     "percent_asian"     "percent_hispanic" 
[7] "per_capita_income" "median_rent"       "median_age"

Exercise: Make a map of some demographic of the counties of your home state. Share it with your neighbor.

My Answer

df_county_demographics$value = df_county_demographics$percent_asian

county_choropleth(df_county_demographics, num_colors=1, state_zoom="new york")

plot of chunk unnamed-chunk-16

Part 3

ZIP Codes

ZIP Code vs. ZCTA

Post Office maintains ZIP Codes
Census maintains Zip Code Tabulated Areas (ZCTA)

ZIP Codes are difficult to analyze:

  • Aren't polygonal
  • Change freqently
  • Span counties and states

ZIP Code Population

library(choroplethrZip)
data(df_pop_zip)
head(df_pop_zip)

    region value
132  01001 17380
133  01002 28718
134  01003 11286
135  01005  5120
136  01007 14593
137  01008  1160

Exercise: What is the population of the zip you are from? Tell your neighbor.

My Answer

df_pop_zip[df_pop_zip$region == "11021", ]

     region value
2811  11021 17449

Boxplot

boxplot(df_pop_zip$value)

plot of chunk unnamed-chunk-19

Exercise: Draw one interpretation from this boxplot

Mapping ZCTAs in a State

zip_choropleth(df_pop_zip, state_zoom="new york")

plot of chunk unnamed-chunk-20

Exercise: Draw one inference from this map

Mapping ZCTAs in a County

# 36059 is the FIPS code for Nassau County, NY
zip_choropleth(df_pop_zip, county_zoom=36059)

plot of chunk unnamed-chunk-21

Exercise: Draw one inference from this map

Exercises

  1. Create a choropleth map showing the population of the zip codes in your home state. Share it with your neighbor.

  2. Create a choropleth map showing the population of the zip codes in your home county. Share it with your neighbor.

More ZIP Demographics

data(df_zip_demographics)
colnames(df_zip_demographics)

[1] "region"            "total_population"  "percent_white"    
[4] "percent_black"     "percent_asian"     "percent_hispanic" 
[7] "per_capita_income" "median_rent"       "median_age"

Exercise: Make a map of some demographic of the ZIPs in your home state. Share it with your neighbor.

My Answer

df_zip_demographics$value = df_zip_demographics$per_capita

zip_choropleth(df_zip_demographics, state_zoom="new york")

plot of chunk unnamed-chunk-23

Exercise: Draw one inference from this map

Exercise

Exercise: Make a map of some demographic of the ZIPs in your home county. Share it with your neighbor.

My Answer

df_zip_demographics$value = df_zip_demographics$per_capita

zip_choropleth(df_zip_demographics, county_zoom=36059, num_colors=1)

plot of chunk unnamed-chunk-24

Exercise: Draw one inference from this map

Part 4

Historic Data

Technical Details - R

library(acs)
# api.key.install("<key>")

Technical Details - Census Bureau

  • Google “CRAN choroplethr” -> click “Mapping US Census Data”

survey-screenshot

New York in 2010

  • ?get_state_demographics 
library(choroplethr)
df_2010 = get_state_demographics(2010, 5)
df_2010[df_2010$region == "new york", 
                          "total_population"]

[1] 19229752

Exercise: What was the population of your home state according to the 2010 5-year ACS?

Part 5

More Data

What's Available?

  • Google “CRAN choroplethr” -> “Mapping US Census Data”

table-list

Example: Public Assistance Income

  • Search for “B19067”: AGGREGATE PUBLIC ASSISTANCE INCOME IN THE PAST 12 MONTHS (IN 2012 INFLATION-ADJUSTED DOLLARS) FOR HOUSEHOLDS 
# 6075 is the county FIPS code for San Francisco
zip_choropleth_acs("B19067", county_zoom=6075, num_colors=3)

plot of chunk unnamed-chunk-27

Wrapping Up

  1. Please provide feedback!

  2. Need a software engineer or data analyst?

    • Want to continue with open source

  3. Keep in touch!